> 4. Win32. I don't know that it's broken, but I'm assuming it is on
> general principle.
D:\build\parrot>perl Configure.pl
[snip]
Probing for C headers...done.
Determining some sizes...Linker failed (see test.ldo) at
lib/Parrot/Configure/St
ep.pm line 170.
The code
# 'link' needs to be link.exe, not cl.exe.
# This makes 'link' and 'ld' the same.
Configure::Data->set('link', Configure::Data->get('ld'));
in config\init\hints\mswin32.pl sets link to a ''", since
D:\>perl -V:ld
ld='';
D:\>
I already reported about that problem a while ago and nobody
responded to it.
This patch
--- parrot/config/init/hints/mswin32.pl 2003-08-20 13:34:27.000000000 +0300
+++ build/parrot/config/init/hints/mswin32.pl 2003-09-15 13:55:22.000000000
+0300
@@ -53,7 +53,8 @@
);
# 'link' needs to be link.exe, not cl.exe.
# This makes 'link' and 'ld' the same.
- Configure::Data->set('link', Configure::Data->get('ld'));
+ Configure::Data->set('link',
+ Configure::Data->get('ld') ? Configure::Data->get('ld') :
'link');
}
if( $is_bcc ) {
Configure::Data->set(
would be okay.
Next.
[snip]
Generating feature.h...done.
Writing Parrot::Config module...done.
Generating Makefiles...value for 'ccwarn' in config/gen/makefiles/root.in is
und
ef at lib/Parrot/Configure/Step.pm line 145, <IN> line 134.
value for 'ccwarn' in config/gen/makefiles/classes.in is undef at
lib/Parrot/Con
figure/Step.pm line 145, <IN> line 14.
value for 'ccwarn' in config/gen/makefiles/imcc.in is undef at
lib/Parrot/Config
ure/Step.pm line 145, <IN> line 26.
done.
D:\build\parrot>nmake
[snip]
io/io_win32.c
io_win32.c
io/io_win32.c(212) : error C2371: 'PIO_win32_flush' : redefinition;
different ba
sic types
io/io_win32.c(49) : see declaration of 'PIO_win32_flush'
NMAKE : fatal error U1077:
'D:\Perl\5.6.1\bin\MSWin32-x86-multi-thread\perl.exe'
: return code '0x2'
Stop.
Seems like Juergen did apply only one of two patches which I proposed on
06.09.2003
See Re: Win32 Build Issues (was Re: Linking pdump and dissassemble)
06.09.2003 4:40
And finally,
D:\build\parrot>nmake test
[snip]
t/pmc/intlist..........ok
t/pmc/io...............NOK 3# Failed test (t/pmc/io.t at line 38)
# got: 'fdopen failed
# '
# expected: 'ok
# '
t/pmc/io...............NOK 4# Failed test (t/pmc/io.t at line 52)
# got: 'fdopen failed
# '
# expected: 'ok
# '
t/pmc/io...............ok 19/19# Looks like you failed 2 tests of 19.
t/pmc/io...............dubious
Test returned status 2 (wstat 512, 0x200)
DIED. FAILED tests 3-4
Failed 2/19 tests, 89.47% okay (less 1 skipped test: 16 okay,
84.21%)
t/pmc/iter.............ok
[snip]
t/pmc/timer............ok
t/native_pbc/number....ok
3/3 skipped: core ops changes
Failed Test Stat Wstat Total Fail Failed List of Failed
----------------------------------------------------------------------------
---
t/pmc/io.t 2 512 19 2 10.53% 3-4
23 subtests skipped.
Failed 1/57 test scripts, 98.25% okay. 2/897 subtests failed, 99.78% okay.
NMAKE : fatal error U1077:
'D:\Perl\5.6.1\bin\MSWin32-x86-multi-thread\perl.exe'
: return code '0xff'
Stop.
So we have only 2 test failing on Windows. Those failings are concerned with
the erroneous implementation of fdopen and Juergen will fix that in a while.
See
Sent: Friday, September 05, 2003 6:24 PM
Subject: [RfC] Semantics of fdopen.
